PROJECT SUMMARY/ABSTRACT/RELEVANCE TO ADRD Abnormalities in speech production in Alzheimer's disease (AD) have received scant attention in the literature. Yet the network architecture of speech motor control circuit ? with its anatomic distribution involving the superior temporal, posterior parietal, premotor, and prefrontal regions ? is indeed a highly vulnerable target in Alzheimer's disease. Determining neural mechanisms of impairments in speech motor control could therefore provide useful scales of network integrity to gauge disease progression and therapeutic efficacy in clinical trials of AD. In our own prior studies in AD (funded by the Alzheimer's Association) we have found that abnormalities in speech motor control clearly exist and can be seen in how patients with AD (ADs) respond to perturbations of pitch in the auditory feedback they hear as they speak. Behaviorally, when pitch feedback is perturbed, ADs respond with significantly larger compensatory responses than controls. Neurally, ADs show greater posterior temporal lobe (pTL) activity and smaller activity in medial prefrontal cortex (mPFC) in response to altered pitch feedback, compared to controls, and these activity changes are correlated with degree of compensation. Both degree of compensation and mPFC activity during compensation are also correlated with measures of cognitive abilities in the ADs, particularly measures of executive function. Here, we propose to investigate the nature of these AD speech motor control abnormalities within the scope of our original funded parent grant. In the parent grant, we develop a computational model of speech motor control and test its predictions using experiments based on magnetoencephalographic imaging (MEGI). In this administrative supplement, we will use these same components of the parent grant to investigate speech abnormalities in AD. We will use the computational model we develop in the parent grant to simulate and mechanistically explain AD speech abnormalities. We will also use a simplified version of the first experiment proposed in the parent grant ? an experiment based on MEGI ? to test hypotheses suggested by the model about the underlying cause of the AD speech abnormalities. The goal of this proposed work is to have a model that can accurately reproduce what goes wrong in AD speech. Such a model ? the first of its kind ? would give us a powerful basis for making many predictions about speech in AD, including predictions about the effects of different patterns of disease progression on speech in AD, which would also allow us to predict patients' responses to different treatments.